Dynamic Routing - PowerPoint Presentation

Slide1

Dynamic Routing

Distance Vector and Link State

RIP

OSPFSlide2

Internet Routing

IP implements datagram forwarding

Both hosts and routers

Have an IP module

Forward datagrams

IP forwarding is table-driven

Table known as routing tableSlide3

Routing Tables

Static routing

Fixes routes at boot time

Requires human intervention

Useful only for simplest cases

Dynamic routing

Table initialized at boot time

Routers communicate to learn new information and update their routing tables continuously:

Protocols

used for information exchange to propagate route data

Data inserted/updated by protocols

automatically

- necessary in large internetsSlide4

4

Autonomous Systems

An

autonomous system (AS)

is a region of the Internet that is administered by a single entity and that has a unified routing policy

Each autonomous system is assigned an Autonomous System Number (

ASN

). Each ASN is either 16bits or 32bits

ASN assigned by Regional Internet Registries

Some are reserved for private use and never appear on the Internet

Example ASNs

U of T

’

s campus network (AS239)

Sprint (AS1239, AS1240, AS 6211, …)Slide5

Number of Autonomous Systems

5Slide6

6

Interdomain

and

Intradomain

Routing

Routing protocols used

inside

an AS, referred to as

intradomain

routing, are called interior gateway protocols (IGP)

Objective: shortest path, only operate within an AS

Routing protocols used

between

ASs, referred to as

interdomain routing, are called exterior gateway protocols (EGP)Objective: satisfy policy of the ASs, not always shortest pathSlide7

7

Interdomain

and

Intradomain

Routing

Intradomain

Routing

Routing within an Autonomous System (AS)

Ignores the Internet outside the AS

Protocols for

Intradomain

routing are collectively called

Interior Gateway Protocols

or

IGP’s. Popular protocols are: RIP (simple, old)OSPF (better) Interdomain Routing

Routing between AS’sAssumes that the Internet consists of a collection of interconnected AS’s

Normally, there is one dedicated router in each AS that handles interdomain traffic.Protocols are collectively called Exterior Gateway Protocols or EGP’

s.Popular protocols are:Border Gateway Protocol (BGP) v4 currentSlide8

8

EGP and IGP

Interior Gateway Protocol (IGP)

Routing is done based on

metrics

Routing domain is

one AS

Exterior Gateway Protocol (EGP)

Routing is done based on

policies

Routing domain is the

entire Internet

1Slide9

9

Components of a Routing Algorithm

A procedure for

sending and receiving

reachability information about a network to other routers

A procedure for

calculating optimal

routes

Routes are calculated using a shortest path algorithm (least “

cost

”)

A procedure for

reacting to and advertising

topology changesSlide10

10

Two Basic

Shortest Path

Routing Algorithms used for

IGP

Distance Vector Routing

Each node knows the distance (cost) to its directly connected neighbors

A node sends periodically a list of routing updates to its neighbors.

If all nodes update their distances to destinations using neighbor information, the routing tables eventually converge

New nodes advertise themselves to their neighbors.

Link State Routing

Each node knows the distance (cost) to its directly connected neighbors

The distance information is broadcast to all nodes in the networkEach node calculates the routing tables independently using global information.Slide11

Summary of DifferencesSlide12

12

IGP Routing Algorithms

Distance Vector

Routing Information Protocol (RIP)

Gateway-to-Gateway Protocol (GGP)

Exterior Gateway Protocol (EGP)

Interior Gateway Routing Protocol (IGRP)

Link State

Intermediate System - Intermediate System (IS-IS)

Open Shortest Path First (OSPF)Slide13

Distance Vector Routing

Initialize routing table with one entry for each

directly

connected Network

Periodically run a distance-vector update to exchange information with routers that are reachable over

directly

connected networksInformation shared is each router’s individual view of network.Uses Bellman-Ford Algorithm to calculate routing table (min cost to all destinations).Slide14

Distance Vector Dynamic Updates

Every router sends a list of its routes to all its neighbors

List contains pairs of:

destination network

,

distance

Receiver replaces/updates entries in its routing table if routing through a neighbor costs less than the current route in its tableReceiver propagates new routes and updates next time it sends out an updateUpdate Algorithm has well-known shortcomings (we will see an example later)Slide15

Update from

a Neighboring Router

Existing Routing Table in a Router “K”

Update received from a neighboring Router “J”.

Net 4 has a better cost via “J”

Net 21 is a new entry learnt from “J”

Net 42, via “J”, has a changed costSlide16

Example of Distance Vector

Assume

:

link cost is 1 on all hops

all updates occur simultaneously

initially each router only knows its directly connected interfaces -> cost = 0

Slide17

Rip Convergence ExampleSlide18

After First UpdateSlide19

After Second UpdateSlide20

After Third UpdateSlide21

Last Update for

ConvergenceSlide22

22

A “

Down

” link has a cost of

Infinity

1

1

Network 4.0.0.0 goes down

Router C marks in its routing table that Net 4.0.0.0 is down

i.e., cost is now “

infinity

”Slide23

23

Characteristics of Distance Vector Routing

Periodic Updates:

Routers exchange Updates with their neighbors periodically (

fixed interval

). Routes that are not refreshed (i.e., timers reset) when an update comes in, are removed from a routers routing table when the timer expires (could be a few update periods).

Triggered Updates:

If a metric changes on a link (usually when a

link goes down

), a router immediately sends out an update for that route without waiting for the end of the update period.\

Full Routing Table Update

:

Most distance vector routing protocols send their neighbors the entire routing table (not only entries which changed).

Route invalidation timers: Routing table entries are invalid if they are not refreshed. A typical value is to invalidate an entry if no update is received for 3-6 update periods.Slide24

Router Routing Table with TimersSlide25

Convergence and Loops

Distance Vector Protocols are subject to

loop formations

because of the

myopic view

of each router.

Routers only hear from neighbors and use that to create a global connectivity map. When changes occur, they are broadcast but take a while to propagate and during that time cycles can form.One particular problem is the count to infinity problem, where updates bounce back and forth and the distance or cost creeps up in value.To counter that, a maximum value is set that once it is reached, the destination is considered to be unreachable and the route is removed from the routing table.Slide26

Down Link and Update from B to C occurs

C notices that NET 4.0.0.0 is down, C removes that entry from its routing table.

C receives a periodic update from B

C sees that B is 1 hop away from Net 4.0.0.0 based on B’s update. It calculates its route to NET 4.0.0.0 using

B as next hop

with a cost of “1+1=2”

C updates its routing tableC will send its neighbors (only B in this case) its new updated routing tableSlide27

Node C sends its Update to B

When C sends its update, B sees the change in cost to NET 4.0.0.0 via C

B updates its entry to Net 4.0.0.0 to “1+2=3” as C is marked as next hop to Net 4.0.0.0. B will then share this new update with its neighbors, including C.

C proceeds to update its entry again for Net 4.0.0.0 (3+1=4) and shares it with B.

B and C repeat this

cycle and the cost increases in value.

Note that B sends updates to A, and its cost for NET 4.0.0.0 will increase correspondingly.Slide28

28

Count-to-Infinity Phenomenon

Why does the count-to-infinity problem occur?

Because each router ONLY has a

“

next-hop-view

”

For example, in the first step, C did not realize that B’s route (with cost 1) to network 4.0.0.0 went through itself and B did not realize that C’s update was based on B’s connectivity information.

How can the Count-to-Infinity problem be solved?

A router with a down link:

Sets a

max value

for the cost. Usually

16 is used to signify infinity.Advertises link with a cost of 16 (triggered update).Any destination with route cost = 16 is considered unreachable and destination is removed from routing table (after triggered update, no longer advertised).Slide29

How to Prevent Count to Infinity

Enhancements proposed to prevent the Count to Infinity problem and routing loops:

Split Horizon

Route Poisoning

Reverse Poison

Hold Down TimersSlide30

Split Horizon

A router

never

sends information about a route in the

direction from which the original information came

. Routers keep track of which neighbor sent information about a route in its routing table. Updates to that route are never sent to that neighbor, unless the latest update is caused by information from a different neighbor.

Router B never sends Router C updates about NET 4.0.0.0 as C is next hop on path to Net 4.0.0.0 When NET 4.0.0.0 goes down, C removes the entry from its table.Updates will no longer include NET 4.0.0.0.B will remove route to NET 4.0.0.0 when the timer expires for that route in its routing table (received no updates for that entry in awhile).Slide31

Route Poisoning and Poison Reverse

Marking a

down

link as a cost of

infinity

.

When NET 4.0.0.0 goes down, router C marks it as “cost = infinity” and advertises the new cost of this network to its neighbors in a TRIGGERED UPDATE.removes the route from its table When B gets C’s update, it:sends a triggered update to all its neighbors with the new cost of infinity for that destination (poison reverse supercedes Split horizon if in use)removes the route from its table If SPLIT Horizon is not being used:B’s update might not get to A before A sends B it’s updates that includes the old information related to the unreachable destination. So:……. we must use SPLIT Horizon and or HOLD DOWN Timers.Slide32

Hold Down Timers

After receiving a route poisoning (cost = infinity) for a route from a neighboring router, a router starts a hold-down timer for that route.

During the hold-down timer, the “downed” route is

marked

.

If the router gets an update from that same neighbor with a

new cost (< infinity) within the hold-down timer period, the hold-down timer is removed and the table is updated (route no longer marked). However, if within the hold-down timer, an update is received for that marked route from another router with a better cost, that update is ignored. In our example, when router B receives a route poisoning update from router C:It marks NET 4.0.0.0 as “down” in its routing table and starts the hold-down timer for NET 4.0.0.0. In this period, if it receives an update from C informing that NET 4.0.0.0 is recovered then B will accept that information, remove the hold-down timer and reinstitute that destination in its routing table. But if B receives an update from A informing it that NET 4.0.0.0 can be reached in X hops (X < infinity), that update will be ignored. When the hold-down timer expires a new update for that route from a neighbor will put it back in router B’s table.Slide33

Poison Reverse

Poison

Reverse - Breaking the Split Horizon rule for updates with cost =

infinity

It basically says that when a router receives a NET is down update from a neighbor (cost = infinity), the router breaks the split horizon rule and sends a triggered route update to all its neighbors including

the originating neighbor with a cost = infinity for that very same destination.For example, when router B receives a route down (i.e., a cost = infinity) for NET 4.0.0.0 from router C then router B will send an update to all its neighbors including router C (which breaks the split horizon rule) with the same cost = infinity for NET 4.0.0.0. Every router performs poison reverse when learning about a down network/link. Slide34

34

RIP - Routing Information Protocol

A simple

intradomain

protocol (Interior Gateway Protocol IGP)

Straightforward implementation of Distance Vector Routing

Each router advertises its distance vector every 30 seconds (or whenever its routing table changes) to all of its neighbors (destination address, distance)

Uses metric of hop count and uses 1 for every hop (link)

Maximum hop count is 15, with

“

16

”

equal to

“

”Routes are timed out (set to 16) after 3 minutes if they are not updatedUses split horizon and poison reverse techniques to solve ``count to infinity and looping’’Current standard is RIPv2Slide35

Two Forms of RIP

Active

Used by routers

Broadcasts routing updates periodically

Uses incoming messages to update routes

Passive

Used by “non forwarding” hostsUses incoming update messages to change route table – changes overwrite ICMP redirectsDoes not send updatesSlide36

RIPv2

Route Update includes

subnet mask

Authentication supported

Explicit next-hop information

Messages are multicast

IP multicast address for RIP is 224.0.0.9Slide37

RIPv2 Update Packet

Route Tag: Used to carry information from other routing protocols (e.g., autonomous system number)Slide38

Description of Fields

Command

 - Indicates whether the packet is a

request

or a

response

. request asks that a router send all or a part of its routing table. response can be an unsolicited regular routing update or a reply to a request. Responses contain routing table entries. Multiple RIP packets are used to convey information from large routing tables.Version - Specifies the RIP version used. For RIP 2 this value is set to 2.Unused - Has a value set to zero.Address-family identifier (AFI) - Specifies the address family used. RIP is designed to carry routing information for several different protocols. Each entry has an address-family identifier to indicate the type of address being specified. The AFI for IP is 2. If AFI for the first entry in the message is 0xFFFF, the remainder of the entry contains authentication information. Route tag - Provides a method for distinguishing between internal routes (learned by RIP) and external routes (learned from other protocols).IP address - Specifies the IP address for the entry.Subnet mask - Contains the subnet mask for the entry. If this field is zero, no subnet mask has been specified for the entry.Next hop - Indicates the IP address of the next hop to which packets for the entry will be forwarded.Metric - Indicates how many internetwork hops (routers) will be traversed in the trip to the destination. This value is between 1 and 15 for a valid route, or 16 for an unreachable route.Slide39

Contd.

Up to 25 routing table entries can be listed in a single RIP packet. If the AFI specifies an authenticated message, only 24 routing table entries can be specified.

RIP has numerous

stability features

:

By placing a

finite limit on the number of hops that a route can take, routing loops are discouraged, if not completely eliminated. Various timing mechanisms that help ensure that the routing table contains only valid routes:The timeout timer is used to help purge invalid routes from a RIP node. Routes that aren't refreshed for a given period of time are likely invalid because of some change in the network. Thus, RIP maintains a timeout timer for each known route. When a route's timeout timer expires, the route is marked invalid but is retained in the table until the route-flush timer expires.Split horizon, poison reverse and hold-down mechanisms that prevent incorrect routing information from being disseminated throughout the network.Slide40

40

RIP Message Exchange

Uses UDP transport

Dedicated port for RIP is UDP port 520

Two types of command messages:

Request messages

used to ask neighboring nodes for an update

Response messages

contains an updateSlide41

41

Routing with RIP

Initialization:

Send a

request packet

on all interfaces requesting routing tables from neighboring routers:

RIPv2 uses multicast address 224.0.0.9

Request received

: Routers that receive above request send their entire routing table

Response received

: Update the routing table

Regular routing updates

: Every 30 seconds, send all or part of the routing tables to every neighbor in a

response

messageTriggered Updates: Whenever the metric for a route changes, send updated route. Slide42

RIP Summary

Slow convergence

Low overhead

Limited to 15 hops (max cost, i.e., infinity =16)

Only uses local information from immediate neighbors for routing decisions - relies on propagation of information for global view of network